U.S. patent number 7,096,055 [Application Number 09/103,533] was granted by the patent office on 2006-08-22 for method to control delivery of radiation therapy.
Invention is credited to Achim Schweikard.
United States Patent |
7,096,055 |
Schweikard |
August 22, 2006 |
Method to control delivery of radiation therapy
Abstract
A method is disclosed for controlling the delivery of radiation
therapy to a tumor of a patient from one or more beams of ionizing
radiation, to conform to a prescribed dosage of radiation for each
of predetermined plural respectively shaped portions of the tumor
according to the shape and other characteristics of the tumor. A
radiation beam is selectively generated from different directions
with respect to travel of the beam along a plurality of spatial
paths including oscillating and arcuate movements. Parameters of
the beam are calculated from conditions of distribution of a target
dose, and the cross-section of the beam is adjusted so as to
deliver the prescribed dosage of radiation to each of the
respectively shaped portions of the tumor on which the beam
impinges. The cross-section of the beam is constantly adjusted
according to a predetermined area of the tumor which is to receive
radiation therapy, and adjusted cross-sections of the beam are
moved along selected individual ones of the spatial paths at least
one time. Also, movements of the beam along individual spatial
paths are split according to the plural portions of the tumor which
are to receive different radiation doses. Travel of the beam is
controlled along the selected individual paths so as to deliver
radiation therapy within the prescribed dosage to each of the
plural portions of the tumor in a minimum amount of time. In one
embodiment, a micro multi-leaf collimator is placed between the
beam and the tumor, and the collimator leaves are adjusted to
change the cross-section of the beam impinging on a specified
portion of the tumor according to the shape of the specified
portion.
Inventors: |
Schweikard; Achim (Hamburg,
DE) |
Family
ID: |
36821798 |
Appl.
No.: |
09/103,533 |
Filed: |
June 24, 1998 |
Current U.S.
Class: |
600/407; 378/65;
606/130 |
Current CPC
Class: |
A61N
5/1031 (20130101); A61N 5/1042 (20130101); A61N
5/1036 (20130101) |
Current International
Class: |
A61B
5/05 (20060101) |
Field of
Search: |
;600/1-8,407-409
;606/130 ;128/653.1 ;364/413.26,578 ;378/65,152 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Lacyk; John P.
Attorney, Agent or Firm: Blank Rome LLP
Claims
What is claimed is:
1. A radiation therapy process for controlling the delivery of a
beam of radiation of adaptable cross-section from any of a
plurality of different directions to impinge on a preselected tumor
region in a patient's body in a treatment session, said process
comprising measuring the boundaries of said tumor region based on a
selected fixed position of the patient's body; adjusting said
boundaries to compensate for shifting of said tumor region from
movement during patient breathing and other natural movements;
subdividing the area occupied by said tumor region into subareas
based upon differences in radiation dosage to be delivered thereto
from said beam; specifying a maximum radiation dose to be delivered
to each respective subarea in the treatment session; determining
the dose rate applicable for each of said subareas for a plurality
of discrete directions of impingement of said beam along a selected
travel path thereof in a predetermined volume of space onto the
respective subareas of said tumor region according to the
distribution of said differences in radiation dosage for the
respective maximum dose to be delivered thereto; calculating a
minimum interval of time for said treatment session in which to
dispense a radiation dosage not more than said maximum dose for all
said subareas of said tumor region; and controlling the intensity,
cross-section and direction of impingement of said radiation beam
on said subareas, to deliver the respective radiation dosages
throughout said tumor region in substantially said minimum interval
of time.
2. The radiation therapy process of claim 1, wherein the step of
controlling said cross-section of said radiation beam includes
adjusting the positions of leaves of a micro-multileaf collimator
on which said beam impinges before impinging on said tumor
region.
3. A method for controlling the delivery of radiation therapy to a
tumor of a patient from one or more beams of ionizing radiation, to
conform to a prescribed dosage of radiation for each of
predetermined plural respectively shaped portions of the tumor
according to the shape and other characteristics of the tumor, said
method comprising the steps of: generating a radiation beam
selectively from different directions with respect to travel of the
beam along a plurality of spatial paths including oscillating and
arcuate movements, to impinge on said plural portions of said
tumor, calculating parameters of said beam from conditions of
distribution of a target dose, and adjusting the cross-section of
said beam so as to deliver said prescribed dosage of radiation to
each of said predetermined plural respectively shaped portions of
the tumor.
4. The method of claim 3, including the step of constantly
adjusting said cross-section of said beam according to a
predetermined area of aid tumor which is to receive said radiation
therapy.
5. The method of claim 4, including the step of moving adjusted
cross-sections of said beam along selected individual ones of said
spatial paths at least one time.
6. The method of claim 5, including the step of splitting movements
of said beam along individual ones of said spatial paths according
to said predetermined plural respectively shaped portions of said
tumor which are to receive different radiation doses from said
beam.
7. The method of claim 3, including the step of controlling travel
of said beam along said selected ones of said individual paths so
as to deliver said radiation therapy within said prescribed dosage
to each of said predetermined plural respectively shaped portions
of the tumor in substantially a minimum amount of time.
8. The method of claim 4, including the step of placing a
micro-multileaf collimator between said beam and said tumor, and
adjusting leaves of said micro-multileaf collimator to change the
cross-section of said beam impinging on a specified portion of said
tumor according to the shape of said specified portion.
9. A process for treating malignant tumors with a beam of ionizing
radiation, comprising the steps of: controlling said radiation beam
to render it of freely adaptable cross section according to
predetermined conditions of distribution of a target dose to be
delivered to a tumor, so as to control the prescribed dose and
distribution of radiation delivered from said beam based on shape
and differing needs of treatment of various portions of said tumor,
including interposing a radiation-opaque device of adjustable shape
in the path of said beam, and automatically adjusting said shape of
the interposed device to enable delivery of radiation distributed
in increments from a predetermined maximum dosage to a
predetermined minimum dosage according to a prescribed pattern for
said differing needs of treatment of said various portions of the
tumor.
10. The process of claim 9, including adapting said beam for travel
along different spatial paths according to said conditions of
distribution.
11. The process of claim 9, including adapting said beam for travel
in oscillating and arc-like movements according to said conditions
of distribution.
12. The process of claim 9, including continuously adjusting said
cross-section of the beam for projection on various portions of the
tumor according to their respective differing needs of
treatment.
13. The process of claim 12, including moving said adjusted cross
sections of the beam a predetermined number of times along
individual ones of plural paths.
14. The process of claim 13, including splitting individual path
movements are into several parts for delivering predetermined
different radiation doses to said various portions of the
tumor.
15. The process of claim 9, including performing said step of
controlling said radiation beam to continuously adjust its
cross-section in a manner to optimize the time interval of
treatment of the tumor.
16. The process of claim 9, including interactively displaying
boundaries of the tumor on a computer monitor screen.
17. The process of claim 9, including using a micro-multileaf
collimator as said device.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention pertains generally to a method for
controlling delivery of radiation beams, and, in particular, to a
method of controlling delivery of radiation therapy for medical
purposes.
2. Description of the Related Art
Radiation therapy is a procedure frequently used in medicine, such
as for shrinkage of tumors. Most small tumors are irradiated with
high-energy photons. The radiation dose for each tumor and patient
must be determined individually. In doing this one must determine
from which direction and with which dose weight one is to give
radiation treatment, given a known total dose that is to be
dispensed.
An especially effective type of radiation therapy is the so-called
pendulum irradiation, in which the source of rays traverses along a
circular path in a space.
Today, planning the radiation procedure is usually performed
manually by the treating physician, who determines both the
direction of the single beam and the dose weight. The cross section
of the beam used is generally rectangular or circular due to the
use of motor-driven leaden jaws or round collimators.
It is, however, necessary to coordinate the shape of the beam with
that of the tumor so that, with optimal and exact tumor radiation
therapy, the tissue surrounding the tumor and/or the healthy organs
are protected against significant exposure to ionizing
radiation.
Toward that end, machines have been developed--so-called
micro-multileaf collimators (MMLC's)--with which as many different
field configurations as desired can be produced. This is
accomplished by bringing into the beam path movable, leaf-like
lamellae that are independent of each other.
Micro-multileaf collimators are also used with pendulum
irradiation. At present it is customary to have up to four circular
paths or oscillating motions in the traversed space. Raising the
number of circular paths is, however, of crucial significance for
increasing the effectiveness of the radiation therapy. Such an
increase in the number of circular paths necessitates a
considerable increase in the time and complexity of planning the
radiation procedure. In part this is because the number of possible
path combinations rises exponentially, requiring that the number of
possible positions of the lamellae of the MMLC must also be
considered. If an MMLC is also used for oscillating motions, the
path of each lamella (for example, the number of lamellae may be
52) must also be calculated. It will be apparent, then, that the
corresponding planning process is much too involved to rely on
manual calculations.
U.S. Pat. No. 5,458,125 of the applicant herein discloses a process
for the partial automatic calculation of the direction of the
single beam and the weights for static radiation (i.e., no
oscillating motion). This planning procedure is only suitable for
conical collimators, i.e., for a circular cross section of the
beam. The applicant herein has also described a procedure in which
only static directions of the single beam are considered, that was
published in the conference volume generated by the CAR Conference
97 (held 25-28 Jun. 1997 in Berlin, Germany). Additional prior art
will be found in U.S. Pat. Nos. 3,987,281 and 4,868,843.
Planning that takes into consideration dose levels can no longer be
carried out manually due to the many possible combinations when
using MMLC with pendulum irradiation.
SUMMARY OF THE INVENTION
It is therefore a principal object of the present invention to
provide a process that enables a determination of whether the
necessary conditions have been fulfilled and to then produce a
rapid and accurate calculation of the path for use with MMLC's. Not
only circular paths, but also the paths of individual leaves of the
MMLC are calculated from the data on the anatomy of the
patient.
Moreover, the process according to the invention may take into
consideration different dose levels. For example, using this
process it is possible to apply radiation therapy accurately to
spots to be irradiated when using a very small target dose. This
goal is significant in connection with the use of MMLC's, such as
to take into consideration in planning the radiation procedure
those changes in location of the target (relocatability)
attributable to breathing or changes in position of other organs of
the patient. The process of the invention may also be used in
static situations. Tolerance limits may also be incorporated into
the planning mechanism.
According to the invention, a method is provided for controlling an
irradiator, in particular, an irradiator utilized for medical
purposes such as for cancer therapy, wherein a beam of freely
adaptable cross section is produced from different directions, and
parameters of the beam are partly or completely automatically
calculated from the conditions of a distribution of a target
dose.
In a first step of the method of the present invention, tumor
boundaries and/or motion tolerances and directions of possible
shifting are fed interactively onto a screen by the physician, by
previous preparation in which a computer-assisted three-dimensional
reconstruction of the tumor region from computerized tomography
(CT) and magnetic resonance (MR) data takes place. In the same step
the physician can break down the target volume and sub-range steps
of different target doses and specify the dose limits.
These specifications are investigated to determine whether they
have fulfilled the necessary conditions in accordance with the
invention. If the conditions are not fulfilled, the physician is
called upon to coordinate the values. Alternatively, these values
can be produced partly or completely automatically.
The paths of motion are subdivided into discrete, definite
intermediate positions. When using a circular path, the path can be
subdivided into 10.degree. steps, with a non-equidistant truncation
of the path also being possible.
According to the invention, every discrete direction of a single
beam is assigned a variable that describes the dose weight of the
direction of the single beam.
The target dose is described by a set of points in space in a
target volume and its environment. It can also be described by the
specification of the upper/lower limits or by the supplementary
conditions of maximizing/minimizing in subregions. Iso-dose regions
can also be defined.
If, for example, p is a point in space in a tumor area that is to
receive a pre-specified minimum target dose a, then there results a
requirement that the dose in p must be larger or equal to a. As a
consequence, a series of points according to the invention
furnishes a series of equations and inequalities that describe the
ordering of the truncated directions of the single beams with the
target volume and target dose. When a point p lies in k single
beams s.sub.1, . . . , s.sub.k and the maximum value of the target
dose is a, one obtains the following inequality: x.sub.1+x.sub.2+ .
. . +x.sub.k.ltoreq.a, where x.sub.1, x.sub.2, . . . , x.sub.k
represent the dose weights for each direction of the single
beam.
In total, one obtains n equalities and inequalities, in which n is
the number of points considered. Whether the system of equations
fulfills the necessary conditions is checked by using mathematical
methods, in which processes of linear programming as well as
genetic algorithms, neuronal networks, or other self-defined
processes can be used due to its linearity.
The system can always be solved if a very small set of marginal
conditions are required, e.g., if only lower limits for the target
dose are defined.
During the process the distribution can be optimized through other
marginal conditions.
By solving the system of equations, one obtains dose values for the
strength of the dose and the dose weight for each discrete
direction of the single beam for each sub-field and each path. The
number of the individual doses calculated in this way is equal to
the number of the variables, which is, in turn, equal to the
product of the number of paths, the number of sub-fields of the
target volume, and the number of discrete directions of the single
beams.
To reduce the number of variables and be able to carry out the
process, all variables according to the invention that correspond
to the same path and the same sub-field, but different directions
of the single beam, are set equal. As an alternative, the same
variable can be used for a path.
In another step variables with a value close to zero are set equal
to zero and the distribution is calculated again.
After calculating the distribution of the directions of the single
beam and the dose weights, the next step is to calculate the paths
of the leaf movements of the MMLC. The supplementary condition for
this calculation is time minimization.
Alternatively, before calculating the paths of the leaf movements,
the calculated dose distribution can be displayed and modified
according to various criteria. Only after confirmation by the user
are the paths calculated.
The subdivision of the target volume into areas that are to receive
different doses of radiation therapy is important for the paths of
the leaf movements. During the first sweep, the field configuration
is equal to the configuration of the projection of the tumor,
whereby the radiation therapy is the smallest assigned dose. During
the subsequent sweeps, the field configuration is matched with
different individual iso-dose areas so the smallest possible path
sweeps are necessary.
A typical plan with a subdivision of the target region into two
subregions consists of four paths in space. This situation
corresponds to eight variables, whereby each path is used twice
with a different position of the collimator and consequently a
different field configuration.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention is illustrated in greater detail in the text below,
in conjunction with the drawings, in which:
FIG. 1 is a diagrammatic perspective top view of a micro-multileaf
collimator (MMLC);
FIG. 2 is a diagrammatic representation of the way a micro-leaf
collimator functions;
FIG. 3 is a diagrammatic representation of the way a micro-leaf
collimator assigned with different doses functions;
FIGS. 4a and 4b are diagrammatic views illustrating the movement
tolerance of a tumor;
FIG. 5 is a perspective view of a tumor in which the central area
is to receive a lower dose than the outer area;
FIG. 6 is a perspective view of a tumor in which the central area
is to receive a higher dose than the outer area;
FIG. 7 is a diagrammatic view of a circular path with discrete
intermediate positions of a tumor subdivided into two areas
receiving different doses;
FIG. 8 is a perspective view of the tumor of FIG. 5 with a
subdivision suitable for irradiation; and
FIG. 9 is a is a flow chart illustrating an exemplary radiation
therapy process for controlling the delivery of a beam of radiation
of adaptable cross-section from any of a plurality of different
directions to impinge on a preselected tumor region in a patient's
body in a treatment session.
DETAILED DESCRIPTION OF THE PRESENTLY PREFERRED EMBODIMENT AND
METHOD OF THE INVENTION
As is shown in FIG. 1, a micro-multileaf collimator MMLC 1 consists
of several individual leaves or lamellae 2, which can be brought
into the beam path independent of each other through conventional
motor-driven means (not shown). In this way, as many field
configurations as desired can be produced and, in particular, the
field configuration can be matched to the projection of the tumor
shape and the tumor area which is to receive the radiation
therapy.
The way in which MMLC 1 functions will be better understood by
reference to FIG. 2. The leaves 2 are moved in such a way that the
cross section of the beam matches the shape of the tumor T.
If, for example, an inner area of the tumor T is to receive a
different, higher dose by using MMLC 1, as is shown in FIG. 3, the
inner area 3 is first given a radiation dose that corresponds to
the difference between the dose assigned to the inner area and the
dose assigned to the outer area. Finally, by matching the cross
section of the beam to the entire tumor area T again, the entire
tumor receives the remaining dose of radiation therapy--the dose
corresponding to that assigned to the outer area.
According to the process of the invention, as a first step tumor
boundaries and/or movement tolerances and the directions of a
possible shifting are introduced interactively onto a computer
screen by the physician after obtaining a computer-assisted
three-dimensional reconstruction of the tumor region from CT and MR
data. FIG. 4a shows a tumor T and its possible direction of
real-time movement in or on the patient's body--e.g., as a result
of breathing by the patient. In FIG. 4b, a tolerance range is
introduced that is plotted such a way that a movement of the tumor
is taken into consideration without including a neighboring organ 4
that must not be irradiated during radiation therapy.
Further according to the process of the invention, the target
volume of the tumor T is subdivided into subareas of different
target doses, and the dose limits are specified, by the physician.
This is represented diagrammatically in FIGS. 5 and 6. In FIG. 5,
the inner area is to receive a smaller dose of radiation than the
outer area. In contrast, FIG. 6 depicts a situation a larger dose
of radiation therapy is to be delivered to the inner area than to
the outer area. The number of areas of different target doses is
unlimited. In this connection, a grid-like subdivision of the tumor
is possible.
According to another aspect of the invention, this information can
be examined to determine whether it fulfills the necessary
conditions. If the conditions are not fulfilled, the doctor is
called upon to adjust the respective values so they match.
Alternatively, these values can be produced partly or wholly
automatically.
Referring to FIG. 7, the movement paths are subdivided into
discrete, definite intermediate positions. In the case of a
circular path, for instance, the path can be subdivided into
10.degree. steps, whereby a non-equidistant truncation of the path
is also possible.
A further aspect of the invention calls for each discrete direction
of the single beam to be assigned a variable value that describes
the dose weight of that direction of the single beam. Finally,
within the framework of the process according to the invention the
variables are assigned values that satisfy the conditions of the
target dose distribution. Weights and directions of the single beam
are calculated first, and after they are determined, the optimal
movements of the lamellae are calculated.
The target dose is described by a set of points in space in the
target volume and its environment. It can also be described by
specifying the upper/lower limits or the supplementary
conditions--i.e., maximization/minimization in subregions.
When a point p of the tumor area lies in k single beams s.sub.1, .
. . S.sub.k, and the target dose has a maximum value of a, one
obtains the following inequality: x.sub.1+x.sub.2+ . . .
+x.sub.k.ltoreq.a, where x.sub.1, x.sub.2, . . . , x.sub.k
represents the dose weight for each direction of the single
beam.
In total, n equations or inequalities are obtained, where n is the
number of the points under consideration. If the system of
equations is determined by mathematical methods to be sufficient to
fulfill the necessary conditions, the equations are then solved to
obtain dose values for the dose strength and dose weight for each
discrete direction of each sub-field and each path. The number of
single doses calculated in this way is equal to the number of
variables, which is, in turn, equal to the product of the number of
paths, the number of sub-fields on the target volume, and the
number of the discrete directions of the single beam.
After calculation of the distribution of the directions of the
single beam and the dose weight, the paths of the leaf movements of
the MMLC are calculated. A supplementary condition for this
calculation is time minimization of the radiation therapy
procedure.
Different ways of subdividing the target volume and different dose
weight subdivisions lead to different total radiation therapy
times. An important part of the process according to the invention
is the partial or completely automatic calculation of a suitable
subdivision and/or suitable movements of the lamellae, which lead
to a global minimization of the total radiation therapy time.
For this purpose, one must take into consideration that the length
of time of the radiation therapy to dispense dose a in a point p is
directly proportional to a, but the radiation time does not depend
on the size of the field at all or is only negligibly dependent on
it. FIG. 8 shows a tumor in which a central area receives a smaller
dose than the boundary area. In this case a dose of 700 units
(e.g., CGY's) is to be attained in the center and a dose of 900
units in the boundary areas.
One possibility for administering radiation therapy is to divide it
up so the upper subregion A receives a radiation treatment of 900
units, the inner subregion 700 units, and the lower region B 900
units. Due to the proportionability, the total time is 900+700+900
units. For this first example, then, the total time amounts to
2,500 units of time.
If, however, another plan is used, a shorter total time results. If
one first applies radiation therapy to the entire area--i.e., 700
units to the inner, upper, and lower region--and then 200 to the
upper area and finally 200 to the lower region, a total time of
700+200+200=1,100 units of time is yielded. The sequence of the
three steps does not play any role.
According to the invention the process for calculating the lamellae
positions consists in the following steps.
All positions of the collimator in which each series of lamellae
approximates a field boundary are calculated. This leads to a
finite number of collimator positions. A variable is used for each
such position pattern. This variable designates the length of time
that each position pattern receives radiation therapy. If the
required target dose in a point p equals a, a supplementary
condition results, requiring that the total sum of the dose
contributions from all position patterns in which p lies in the
beam must equal a. The sum of all the variables is then minimized.
Because all equations or inequalities are linear, linear
programming methods or other suitable processes can be used. The
result is the ability to calculate planes that are minimized with
respect to time on a global basis, because, for example, linear
programming processes yield a global minimum. For this reason very
complex movement patterns of the collimator lamellae, in which the
target region is made of many subregions of different doses, can
also be determined.
FIG. 9 is a flow chart derived from the foregoing description,
illustrating an exemplary radiation therapy process for controlling
the delivery of a beam of radiation of adaptable cross-section from
any of a plurality of different directions to impinge on a tumor in
a patient's body in a treatment session.
In the process, the boundaries of the tumor are measured based on a
selected fixed position of the patient's body; and thereafter
adjusted to compensate for shifting of the tumor in response to
natural movement such as patient breathing. The area occupied by
the tumor is subdividied into subareas based on differences in
radiation dosage to be delivered to them from the beam, and a
maximum radiation dose is specified for delivery to each subarea in
the treatment session.
The dose rate applicable for each subarea for a plurality of
discrete directions of beam impingement along a selected travel
path in a predetermined volume of space including the tumor is
determined according to the distribution of the differences in
radiation dosage for the respective maximum dose to be delivered to
each subarea. A minimum time interval for the treatment session is
calculated for dispensing a radiation dosage not more than the
maximum dose to be delivered to all subareas of the tumor. And the
intensity, cross-section and direction of impingement of the
radiation beam on the subareas are controlled to deliver the
respective radiation dosages throughout the tumor in substantially
that minimum time interval.
* * * * *